QoS Analysis for Serverless Computing Using Machine Learning

Abstract

Large-scale computing systems are becoming more popular as the need for computing power increases every year. Serverless computing has emerged as a powerful and compelling paradigm for the hosting services and applications because of the rapid shift in business application architectures for containers and microservices. Further, Serverless computing offers economical services and scalability to fulfil the growing demand of computing in a timely manner. Therefore, it is important to analyse the Quality of Service (QoS) of Serverless Computing systems to monitor its performance. In this chapter, we used the latest machine learning models to predict system configurations in Serverless computing environments. Knowing about system configurations in advance helps to maintain the performance of the system by analysing QoS. Further, a no-cost model is proposed to examine and compare different configurations of workstations in serverless computing environments. To achieve this, we deployed Theoretical Moore’s, Fitted Moore’s, 2-D poly regression and 3-D poly regression machine learning models following Graphics Processing Unit (GPU) requirements and compare the results. The experimental results demonstrated that Fitted Moore was the best model with an R2 score of 0.992.